-
Notifications
You must be signed in to change notification settings - Fork 0
/
training_qlora.py
71 lines (62 loc) · 1.83 KB
/
training_qlora.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
import torch
from transformers import (
AutoModelForCausalLM,
AutoTokenizer,
BitsAndBytesConfig
)
from peft import (
LoraConfig,
get_peft_model,
prepare_model_for_kbit_training
)
def load_model_for_training(model_name):
bnb_config = BitsAndBytesConfig(
load_in_4bit = True,
bnb_4bit_use_double_quant=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_compute_dtype=torch.bfloat16,
)
model = AutoModelForCausalLM.from_pretrained(
model_name,
device_map="auto",
trust_remote_code=True,
quantization_config=bnb_config,
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
tokenizer.pad_token = tokenizer.eos_token
model.gradient_checkpointing_enable()
model = prepare_model_for_kbit_training(model)
lora_config = LoraConfig(
r=16,
lora_alpha=32,
target_modules=["query_key_value"],
lora_dropout=0.05,
bias="none",
task_type="CAUSAL_LM"
)
model = get_peft_model(model, lora_config)
return model, tokenizer
def train_model(model, tokenizer, data, OUTPUT_DIR):
training_args = transformers.TrainingArguments(
per_device_train_batch_size=1,
gradient_accumulation_steps=4,
num_train_epochs=1,
learning_rate=2e-4,
fp16=True,
save_total_limit=3,
logging_steps=1,
output_dir=OUTPUT_DIR,
max_steps=80,
optim="paged_adamw_8bit",
lr_scheduler_type="cosine",
warmup_ratio=0.05,
report_to="tensorboard",
)
trainer = transformers.Trainer(
model=model,
train_dataset=data,
args=training_args,
data_collator=transformers.DataCollatorForLanguageModeling(tokenizer, mlm=False),
)
model.config.use_cache = False
trainer.train()